Armin Alaghi, a graduate student in the Computer Science and Engineering program, received a Rackham Predoctoral Fellowship to support his research while he completes his dissertation entitled, Analysis and Synthesis of Efficient Stochastic Circuits. The Rackham Predoctoral Fellowship is awarded to outstanding doctoral candidates in the final stages of their program who are unusually creative, ambitious and risk-taking.
Mr. Alaghi’s current research and dissertation focuses on stochastic computing (SC), which is an unconventional technique that processes data in the form of bit-streams that denote probabilities. Stochastic circuits are small and low-power, but have been seen as impractical because of their accuracy issues and relatively low speed. He believes in the light of the current technology trends, SC’s unique features can be exploited, and SC can become an important alternative to conventional computing.
Alaghi's dissertation begins with a comprehensive survey of SC from a modern perspective, and identifies its key challenges and open problems. He then develops a systematic design methodology for stochastic circuits, by discussing a fundamental relation between SC and spectral transforms. Also, he addresses the accuracy problems by providing a rigorous mathematical framework, and new methods of (pseudo-) random number generation. Finally, he studies several promising SC applications such as image processing, and showed that stochastic implementations can outperform conventional ones.
His research interests include digital system design and testing, approximate computing, and embedded computing.
Armin Alaghi is advised by Prof. John Hayes.
Posted: October 13, 2014